Always good to see more activity in the language learning space. I wish you luck now that you've moved past Anki.
A piece of feedback: one of the common issues I've found with AI-generated questions & answers based on an article is that they will often hone in on testing values. It looks like incontextlearning.com often suffers from this same issue (over half my comprehension questions were "how many"-style). I can easily answer these types of questions even if I don't know what the content is about.
It looks like you can’t do much on incontextlearning.com without logging in. It would be good to have more explanation of what it does and maybe a demo before taking that step.
The Japanese immersion language learning rabbit hole goes very very deep, and it super interesting.
I'm writing some info about it here because I just don't want it to be lost to the sands of the digital times:
It all started way back when with this guy named Khatzumoto who did his own guerilla academic research, for lack for a better word - and proving it by getting fluent in 18 months.
The theories were based in comprehensible input (from Stephen Krashen) and nonstop immersion, which do carry weight and are great ways - if not the only way to learn ANY language.
He created AJATT, All Japanese All The Time, and basically created a mini cult that legitimately got people to fluency.
It might sound extreme but on the other end I heard of someone who did duolingo (100% slop product btw) for 6 months and didn't know how to say "Thank You" in Japanese.
Khatzumoto's ideas were manic, strange, but sometimes truly brilliant. I've never found a blog quite like it ever since - writings that an LLM really can't emulate.
The original blog 404'd but it's been revived by a community member here - https://alljapanesealltheti.me/index.html
I go back to them every now and then when I want some crazy motivation.
Nobody knows what has happened to Khatzumoto, he basically just dropped off the face of the internet - I wonder if he's doing alright.
> The enemy is the static card. It always has the same front, formatting, and font. After enough reps I would latch on to little cues that are irrelevant to the meaning of the card, meaning I would skip the very important step of thinking deeply about the content. A fairly common occurrence was that a word in the sentence would remind me about the meaning of the sentence, giving away the answer to the target word, which is quite different from piecing together the meaning of the target word in a brand new sentence.
Interestingly, that's the "trick" behind a lot of the seemingly magic skill of geo guessers. The best players have played so much, that they now "see" things that a regular person wouldn't even consider to look for, like the camera quality, what year the car was from, and so they narrow down the possible countries by those aspects, before even looking at the "picture".
Hallucinations in LLMs when learning is dangerous; IF you have some background, you can usually tell with LLMs go off the rails, but It would be unfortunate for you to commit to memory an incorrect fact at such a vulnerable time. It will be difficult to "uncommit it" at that point.
Alas, the mote around LLM integration is practically non-existent so I'd think that productization around this would be next to impossible.
Anki is already extremely extendable so I would think that with a not too much work deep LLM integration could be implemented in Anki. Like, instead of showing static content for a card, have Anki call an LLM to create the daily iteration of a given prompt.
Facts are required for critical thinking.
I wonder if the writer eventually transitioned from fact-memorization flashcards, to concept-based questions (how does it make sense that....?).
Spaced repetition is incredibly inefficient. Despite exponential guarantees, there's information and incredibly complex concepts that are with me forever, after learning about it once.
Yet somehow a German word is impossible to remember.
LLMs suck, because their goal is not to improve learning. Same way Duolingo sucks, the goal is to optimize global metrics over a massive userbase, not optimize individual metrics, where each individual has its own context.
I think that the author was using Anki incorrectly, and that led them to the spurious conclusion that "Anki is dead". I also have attempted to use Anki this way- using someone else's deck to try to force myself to learn something new. But that doesn't work, because it is just memorizing random symbols, as they noted in the article ("The enemy is the static card"). For example in maths learning, memorizing arbitrary terms, symbols, etc is useless. However, once I am introduced to a concept for the first time, then I add it to my Anki deck so I can make sure I remember it. The key is the context, and writing terms / definitions etc that speak specifically to me. I still need to work out different variations of the concept to understand it, and that's not something that Anki can help with.
I haven't used Anki for language learning, but I imagine that if I did, it would be to add some new vocabulary I had just learned from a book, conversation, film, etc. I don't think it would help me learn a language from zero though- that would require practicing it.
In summary, Anki is great for reinforcing something you've just learned, but you can't reinforce your way into the context that is necessary to truly understand something.
I use Anki to study all kinds of subjects, and more than half of the value is processing what I learned (usually with pencil and paper) into good anki cards ("atomic", as per Michael Nielsen's definition), including the insights that I had when studying the subject (like "what's the comparison with X that I used to understand Y?")
I'm not sure if it is efficient, mind you, but I suppose it's effective because I can recall information later when relevant, and I believe that like exercising just being able to stick to a study routine ends up being more important than picking the best routine
It helps you discover reading material suited to your current knowledge. It’s better to acquire new words within colorful contexts, and then use flashcards to review them after learning the material. (It has its own flashcard companion app or you can use Anki.)
Soon I am working on making the activity of reading words in native texts also count as reviewing those words in current and future flashcards, using FSRS. So that you can spend more time reading and not see it as detracting from catching up on your flashcard review workload. And because the reader tracks every word and kanji you come across, it can start to find and suggest the most effective passages to revisit or read for the first time from your personal corpus it currently accumulates from what you load in.
But next I am working on adding manga and video to enhance the fun of it, as OP mentions being important too.
I’ve recently managed to go full-time on this project and hope to bring it to more languages and platforms before long.
I found Anki way too heavy and the available decks mostly wrong way around for actually learning vocabulary (that is they were from foreign language to native, instead of the other way around) and switching the direction was way too cumbersome.
I have not yet found a really good tool for learning Mandarin, except for classes and actually talking with people and doing the hard work of writing the characters again and again, for which I rarely have energy or patience.
One thing I did notice in a course was, that writing an article about a topic helped a lot. It needs to be something where you use the same new vocabulary many times. But the problem with that is, that it makes my hands and wrist hurt after a couple of writings.
I used Anki pretty heavily during ~8 years of law school and it was a game-changer. Two things that helped me the most (and that I think apply to any kind of flashcards, digital or paper):
Appropriate use of Anki isn't so simple, it's a skill in itself. Here's my strategy specifically to learn vocabulary. TL = target language, NL = native language.
1- front: image+subtitle (in TL) back: word in TL.
2- fill-in-the-blank phrases for the word, fully in TL, translation shows up after completion (you HAVE to use the function where you actually type it out)
3- front: word in TL back: translation in NL with an image, also the inverse, but the image is always in the back. Making it a different picture as the one for 1 is essential
So each new word would generate me about 6 to 8 new cards. At a fast enough rate of card creation, you won't run into the problem of memorizing each card because you will be creating like 100 cards in a day. The "fatal mistake" (to quote the author) of this article is underestimating how much this process of card creation and organization aids in learning. Creating your own study material IS studying in itself.
That is, assuming the strategy being compared to LLMs here is the correct one of actually studying the language and creating your own Anki deck while you study the material, instead of the incorrect strategy of downloading a deck
This basically matches my experience with Anki exactly. It's too easy to learn something that's on the front of the card other than the piece of information that you are trying to learn. I've also been thinking about the same solution.
>All this set me up well to understand that the future of learning would change when in late 2022 the LLM kicked down the door, tracked mud across the carpet, ate everything in the fridge, and demanded more snacks.
Thanks for the laugh, I like your writing style but to echo others, I think you went a little extreme on the Anki.
Using Anki myself for language learning and can definitely recognize the pitfall of unrelated cues triggering the recall, but so far it's been working quite well. I'm mainly using it to practice output though, so the front side of the card is in English.
On semi-related note, currently making on a language app (gengengo.com) if anyone wants to check it out.
Just in case someone has no idea what Anki is like me:
Anki (US: /ˈɑːŋki/, UK: /ˈæŋki/; Japanese: [aŋki]) is a free and open-source flashcard program. It uses techniques from cognitive science such as active recall testing and spaced repetition to aid the user in memorization.[4][5] The name comes from the Japanese word for "memorization" (暗記).[6]
The SM-2 algorithm, created for SuperMemo in the late 1980s, has historically formed the basis of the spaced repetition methods employed in the program. Anki's implementation of the algorithm has been modified to allow priorities on cards and to show flashcards in order of their urgency. Anki 23.10+ also has a native implementation of the Free Spaced Repetition Scheduler (FSRS) algorithm, which allows for more optimal spacing of card repetitions.[7]
Anki is content-agnostic, and the cards are presented using HTML and may include text, images, sounds, videos,[8] and LaTeX equations. The decks of cards, along with the user's statistics, are stored in the open SQLite format.
38 comments
[ 3.1 ms ] story [ 42.1 ms ] threadA piece of feedback: one of the common issues I've found with AI-generated questions & answers based on an article is that they will often hone in on testing values. It looks like incontextlearning.com often suffers from this same issue (over half my comprehension questions were "how many"-style). I can easily answer these types of questions even if I don't know what the content is about.
It all started way back when with this guy named Khatzumoto who did his own guerilla academic research, for lack for a better word - and proving it by getting fluent in 18 months. The theories were based in comprehensible input (from Stephen Krashen) and nonstop immersion, which do carry weight and are great ways - if not the only way to learn ANY language. He created AJATT, All Japanese All The Time, and basically created a mini cult that legitimately got people to fluency. It might sound extreme but on the other end I heard of someone who did duolingo (100% slop product btw) for 6 months and didn't know how to say "Thank You" in Japanese. Khatzumoto's ideas were manic, strange, but sometimes truly brilliant. I've never found a blog quite like it ever since - writings that an LLM really can't emulate. The original blog 404'd but it's been revived by a community member here - https://alljapanesealltheti.me/index.html I go back to them every now and then when I want some crazy motivation.
Nobody knows what has happened to Khatzumoto, he basically just dropped off the face of the internet - I wonder if he's doing alright.
Interestingly, that's the "trick" behind a lot of the seemingly magic skill of geo guessers. The best players have played so much, that they now "see" things that a regular person wouldn't even consider to look for, like the camera quality, what year the car was from, and so they narrow down the possible countries by those aspects, before even looking at the "picture".
* Confused where in the original series Spock goes to school.
* Watches the video and sees 2009 "Star Trek"
* "as a kid"...
* feels old
Anki is already extremely extendable so I would think that with a not too much work deep LLM integration could be implemented in Anki. Like, instead of showing static content for a card, have Anki call an LLM to create the daily iteration of a given prompt.
Yet somehow a German word is impossible to remember.
LLMs suck, because their goal is not to improve learning. Same way Duolingo sucks, the goal is to optimize global metrics over a massive userbase, not optimize individual metrics, where each individual has its own context.
I haven't used Anki for language learning, but I imagine that if I did, it would be to add some new vocabulary I had just learned from a book, conversation, film, etc. I don't think it would help me learn a language from zero though- that would require practicing it.
In summary, Anki is great for reinforcing something you've just learned, but you can't reinforce your way into the context that is necessary to truly understand something.
I'm not sure if it is efficient, mind you, but I suppose it's effective because I can recall information later when relevant, and I believe that like exercising just being able to stick to a study routine ends up being more important than picking the best routine
https://reader.manabi.io
It helps you discover reading material suited to your current knowledge. It’s better to acquire new words within colorful contexts, and then use flashcards to review them after learning the material. (It has its own flashcard companion app or you can use Anki.)
Soon I am working on making the activity of reading words in native texts also count as reviewing those words in current and future flashcards, using FSRS. So that you can spend more time reading and not see it as detracting from catching up on your flashcard review workload. And because the reader tracks every word and kanji you come across, it can start to find and suggest the most effective passages to revisit or read for the first time from your personal corpus it currently accumulates from what you load in.
But next I am working on adding manga and video to enhance the fun of it, as OP mentions being important too.
I’ve recently managed to go full-time on this project and hope to bring it to more languages and platforms before long.
I have not yet found a really good tool for learning Mandarin, except for classes and actually talking with people and doing the hard work of writing the characters again and again, for which I rarely have energy or patience.
One thing I did notice in a course was, that writing an article about a topic helped a lot. It needs to be something where you use the same new vocabulary many times. But the problem with that is, that it makes my hands and wrist hurt after a couple of writings.
- The twenty rules of formulating knowledge (https://www.supermemo.com/en/blog/twenty-rules-of-formulatin...) ; old but really solid advice on how to actually write cards that stick.
- The book Fluent Forever; it’s meant for languages, but the general principles carry over to learning basically anything.
For example I use LLMs to generate cards for me, and Anki's algorithm to make them stick.
Similarly a LLM plugin could easily present a fresh sentence each time you review a particular vocab
1- front: image+subtitle (in TL) back: word in TL.
2- fill-in-the-blank phrases for the word, fully in TL, translation shows up after completion (you HAVE to use the function where you actually type it out)
3- front: word in TL back: translation in NL with an image, also the inverse, but the image is always in the back. Making it a different picture as the one for 1 is essential
So each new word would generate me about 6 to 8 new cards. At a fast enough rate of card creation, you won't run into the problem of memorizing each card because you will be creating like 100 cards in a day. The "fatal mistake" (to quote the author) of this article is underestimating how much this process of card creation and organization aids in learning. Creating your own study material IS studying in itself.
That is, assuming the strategy being compared to LLMs here is the correct one of actually studying the language and creating your own Anki deck while you study the material, instead of the incorrect strategy of downloading a deck
Thanks for the laugh, I like your writing style but to echo others, I think you went a little extreme on the Anki.
On semi-related note, currently making on a language app (gengengo.com) if anyone wants to check it out.
Anki (US: /ˈɑːŋki/, UK: /ˈæŋki/; Japanese: [aŋki]) is a free and open-source flashcard program. It uses techniques from cognitive science such as active recall testing and spaced repetition to aid the user in memorization.[4][5] The name comes from the Japanese word for "memorization" (暗記).[6]
The SM-2 algorithm, created for SuperMemo in the late 1980s, has historically formed the basis of the spaced repetition methods employed in the program. Anki's implementation of the algorithm has been modified to allow priorities on cards and to show flashcards in order of their urgency. Anki 23.10+ also has a native implementation of the Free Spaced Repetition Scheduler (FSRS) algorithm, which allows for more optimal spacing of card repetitions.[7]
Anki is content-agnostic, and the cards are presented using HTML and may include text, images, sounds, videos,[8] and LaTeX equations. The decks of cards, along with the user's statistics, are stored in the open SQLite format.
https://en.wikipedia.org/wiki/Anki_(software)